“The science of today is the technology of tomorrow.”
— Edward Teller

DeepSeek-R1: pure reinforcement learning (RL), no supervised fine-tuning (SFT), no chain-of-thought (CoT) #1minPapers 🧵👇

Takeaways from JPM Healthcare Conf 2025 #JPM2025 For having survived the past two years of biotech winter and current political uncertainties, the crowd was pretty cautiously optimistic for dealflow to recover. And yes, the word “agentic” AI should have been a drinking game.🧵👇

#1minPapers MSFT’s rStar-Math small language model self-improves and generates own training data - second time in recent months that a small model performed equally well (or better) than the billion-parameter large models. 🧵👇

#1minPapers Francois Chollet: use LLMs for tree-search instead of next token prediction. Not a paper, but 90min of Chollet with MLStreetTalk is always worth watching! 🧵👇

#1minPapers “Fourier Analysis Networks” — image and speech processing and generation at lower costs: 🧵👇

#1minPapers Ability to leverage the tools increases with model parameters — “Toolformer: Language Models Can Teach Themselves to Use Tools” — Schick el al

#1minPapers The scare that o1 was able to duplicate its weights via deception got me interested in how LLMs can role-play. Dug up this 2023 paper from DeepMind, ICL, Eleuther AI: “describe [AI] behavior in high-level terms without falling into the trap of anthropomorphism.” 🧵👇

#1minPapers Got something spicy today: a team at Fudan University in China attempted to reproduce o1. This paper was published 2 weeks ago, and full of useful nuggets. Highlighting 3 here: 🧵👇